Nowadays, there is a large demand for deformable conductive materials in applications such as flexible electronics, EMI shielding or soft robotics. Immiscible blends containing conductive nanoparticles form a versatile class of materials with excellent possibilities to tune electrical functionality by means of tailoring their morphology and particle localisation. In the present work, the conductivity of two-phasic blends consisting of poly–alpha-methyl-styrene-co-acrylonitrile and polymethylmethacrylate (PAMSAN/PMMA) containing multiwalled carbon nanotubes (MWNTs) was optimized by tailoring the blend morphology with a polystyrene–polymethylmethacrylate(PS-PMMA) copolymer. The PAMSAN/PMMA blend has a lower critical solution temperature (LCST) and undergoes spinodal decomposition upon heating above the LCST during which the MWNTs selectively localize in the PAMSAN phase. The effects of the molecular weight, architecture (block or random) and concentration of the copolymer on blend conductivity, linear viscoelastic moduli and morphology were systematically investigated. The copolymer induces a huge conductivity increase, whereby blends with only 0.5 wt% MWNTs and 0.25 wt% copolymer exhibit the same conductivity as blends with 2 wt% percolated MWNTs [1]. The increase in conductivity is caused by a morphology refinement and increased degree of cocontinuity with copolymer, leading to double percolated networks. At low molecular weight a block copolymer has a more pronounced effect whereas at high molecular weight the difference between random and block copolymers vanishes. Finally, the dielectric spectra of the blends show an interfacial relaxation peak whose dielectric strength and relaxation time reflect the contact resistance and interfacial capacitance of PAMSAN entrapped between neighbouring MWNTs. From this microcapacitor behaviour, an estimate of the gap spacing between MWNTs was obtained for the different systems. [1] Bharati et al., Polymer 79, 271-282 (2015